Binding of human hemoglobin by Haemophilus influenzae

Transcription

1 FEMS Microbiology Letters 118 (1994) Federation of European Microbiological Societies /94/$07.00 Published by Elsevier 243 FEMSLE Binding of human hemoglobin by Haemophilus influenzae Margaret E. Frangipane, Daniel J. Morton, James A. Wooten, Judith M. Pozsgay and Terrence L. Stull * Division of Infectious Disease, Department of Pediatrics and Department of Microbiology and Immunology, Medical College of Pennsylvania, 3300 Henry Avenue, Philadelphia, PA 19129, USA (Received 17 February 1994; revision received 1 March 1994; accepted 3 March 1994) Abstract: Binding of biotinylated human hemoglobin to Haemophilus influenzae was detected when organisms were grown in heme-deplete, but not heme-replete, conditions. Hemoglobin binding was completely inhibited by a 100-fold excess of unlabelled human hemoglobin or human hemoglobin complexed with human haptoglobin. Binding was only partially inhibited by rat hemoglobin, bovine hemoglobin, human globin, and bovine globin, and not at all by heme, human serum albumin, bovine serum albumin, human transferrin, or myoglobin. Hemoglobin binding was saturable at ng of hemoglobin per 109 cfu. Binding of human hemoglobin was detected in serotypes a-f and serologicauy non-typable strains of H. influenzae, as well as Haemophilus haemolyticus but not Haemophilus parainfluenzae, Haemophilus aphrophilus, Haemophilus parahaemolyticus, or Escherichia coli. Key words: Haemophilus influenzae; Hemoglobin; Iron uptake; Heme uptake Introduction Iron is an essential nutrient for bacterial growth [1]. There is an abundance of iron in the human body; however, the availability of iron to invading microorganisms is strictly limited [1]. In blood, most iron is in erythrocytic hemoglobin; free hemoglobin, released by lysis of red blood cells (RBCs), is rapidly bound by the serum protein haptoglobin [2]. Ferric iron is sequestered by the host iron-binding glycoproteins transferrin and lactoferrin [2]. Organisms must overcome this * Corresponding author. Tel.: (215) ; Fax: (215) iron limitation in order to establish an infection; thus, iron acquisition is an important step in pathogenesis [1]. Pathogenic organisms have developed specific mechanisms of acquiring iron which are inducible under conditions of iron limitation. Such mechanisms include the production of low molecular mass iron binding compounds, termed siderophores, by a wide range of microbial species, although apparently not Haemophilus influenzae [1,3,4]. A number of microbial species including H. influenzae, the Neisseriaceae, and Bordetella species directly bind host iron-binding glycoproteins [1]. H. influenzae is responsible for many human infections, including otitis media, meningitis, epiglottitis, and pneumonia [5]. As well as requir- SSDI (94)

2 244 ing iron, H. influenzae has an absolute requirement for a porphyrin source in order to grow. Previous studies have investigated potential sources of heme and iron available within the human host for their ability to supply the heme and/or iron requirements of H. influenzae in vitro [6-8]. Transferrin is able to satisfy the iron requirements of 11. influenzae and has been shown to be both utilized by, and bound at the cell surface of iron-deplete and heme-deplete H. influenzae [4,9]. The heme requirement of 11. influenzae can be satisfied by heme, human hemoglobin, human hemoglobin complexed to human haptoglobin, and heme complexed to either human hemopexin or human serum albumin [7]. In addition, hemolysis, which occurs in patients with invasive H. influenzae disease [10], results in the release of erythrocytic hemoglobin. Therefore, we investigated whether H. influenzae is able to bind hemoglobin as a possible initial step in the acquisition of heme from hemoglobin. Materials and Methods Bacterial strains and media Type b and serologically non-typable strains of H. influenzae were kindly provided by J. Musser, Baylor University, Houston, TX. Strains of H. influenzae serotypes a-f were obtained from the American Type Culture Collection. Strains of H. aphrophilus, H. parainfluenzae, H. haemolyticus, and H. parahaemolyticus were kindly provided by J. Mortensen, St. Christopher's Hospital for Children, Philadelphia, PA. Heine-replete growth was performed using brain-heart-infusion (BHI) medium supplemented with 10 /xg m1-1 of both heme and J-NAD. For heme-deplete growth the heme concentration was 0.1 /zg ml -~. E. coli strain HB101 was cultured with LM media; for iron depletion, ethylene diamine-di-(orthohydroxyphenyl) acetic acid (EDDA) was added at 500 tzm. H. haemolyticus was heme-restricted as above. Other Haemophilus species were iron-restricted by the addition of 100 izm EDDA and 10 /xg ml ~ protoporphyrin IX to BHI supplemented with 10/zg m1-1 /3-NAD. Chemicals Human, bovine, and rat hemoglobins, human and bovine globins, human haptoglobin, human serum albumin, bovine serum albumin, human transferrin, sperm whale skeletal muscle myoglobin, and heme were obtained from Sigma (St. Louis, MO). All proteins were % pure as specified by the manufacturer. Biotinylation of hemoglobin Human hemoglobin was biotinylated using a modification of the procedure described by Morton and Williams [4]. Hemoglobin (1 mg m1-1) was dissolved in phosphate-buffered saline (PBS), ph 7.4. NHS-LC biotin (Pierce, Rockford, IL) was dissolved to a concentration of 2 mg ml-1 in water at 50 C and then diluted 1 : 10 in the stock hemoglobin solution. After incubation for 2 h at room temperature unbound biotin was removed by passage through a Sephadex G-15 column (Pharmacia, Piscataway, N J). Dot blot assay Bacteria were tested for human hemoglobin binding using a dot blot assay. Bacteria from overnight culture were washed and resuspended in PBS to 107 cfu ml-1; 100 tzl was filtered onto nitrocellulose membranes in a dot blot manifold. Membranes were incubated in 1% skim milk in PBS for 1 h. The organisms were then probed with 200 /zl (500 ng m1-1 stock) of biotinylated hemoglobin per well for 1 h. The membrane was washed three times in 1% skim milk in PBS, incubated with 30 /zl per well of streptavidin conjugated to horseradish peroxidase (HRP) (Accurate Chemical Co., Westbury, NY) for 20 min, and washed three times in PBS. All steps were performed at room temperature. Binding of the hemoglobin was detected by development with 4-chloro-l-naphthol as previously described [9]. In some assays, whole cells were incubated with 1 mg ml-1 trypsin at 37 C for 45 min followed by trypsin inhibitor for 5 min and extensive washing in PBS prior to probing. Whole cell enzyme-linked hemoglobin binding assay Saturability of hemoglobin binding was determined as follows. Heme-deplete H. influenzae

3 245 were washed with cold PBS (ph 7.4) and resuspended to 109 cfu ml- t, in 200 ~l of biotinylated hemoglobin in 1% skim milk; the concentration of hemoglobin ranged from 0 to 200 ng m1-1 (0-3.1 nm). Following incubation for 30 rain, cells were washed twice in cold PBS, and resuspended in 100/~I of streptavidin-peroxidase conjugate (100 ng ml-1). After incubation for 20 min, cells were washed twice in cold PBS, resuspended in 100 /zl of the chromogenic substrate ABTS (1-Step TM ABTS; Pierce, Rockford, IL), and incubated for 30 min. Cells were removed by centrifrugation and supernatants placed in a microtiter plate; 50 ~1 of 1% SDS was added to each well to stop the reaction, and the absorbance was read on a BioTek EL-308 EIA reader at 405 nm. The reaction containing no hemoglobin served as the blank for the assay. All incubations were at 4 C. Quantitative cultures were performed to determine the number of viable cells used in the assay. Isolation of outer membrane proteins Outer membrane proteins were isolated by selective solubilization with Triton X-100 according to the procedure previously described [11]. Results and Discussion H. influenzae type b strain HI689 was grown in heme-replete or heme-deplete conditions and binding of biotinylated human hemoglobin assessed in a whole-cell dot-blot assay (Fig. 1). Cells grown in heme-deplete conditions bound biotinylated hemoglobin (A1); however, cells grown in heme-replete conditions did not (A2). Heme-deplete cells probed with unlabeled human hemoglobin yielded negative results (A4), as did unprobed heme-deplete cells (A3). Nonspecific binding of either the biotinylated hemoglobin or streptavidin-hrp conjugate to the nitrocellulose was not detected (A5). To demonstrate the specificity of binding, competition assays were performed by incubating heme-deplete HI689 with biotinylated human hemoglobin in the presence of a 100-fold excess of the competing compound. Binding of biotiny- A Fig. 1. Dot blot assay demonstrating binding of Hb B to heine-deplete cells. Each well contains approximately 106 cfu/well. Well AI: heme-starved cells, A2: heine-replete cells, A3: heme-deplete cells with no Hb B, A4: unlabeled Hb, AS: no cells, A6: 100-fold excess of unlabeled human Hb, BI: 100-fold excess of rat Hb, B2: 100-fold excess of bovine Hb, B3: 100-fold excess of Hb-Hp complex, B4: 100-fold excess of HSA, B5: 100-fold excess of BSA, and B6: 100-fold excess of Tf. Hb B, biotinylated human hemoglobin; Hb-Hp, human hemoglobin complexed to haptoglobin; HSA, human serum albumin; BSA, bovine serum albumin; Tf, human transferrin. lated hemoglobin was completely inhibited in the presence of excess unlabelled human hemoglobin (A6). When biotinylated hemoglobin was mixed with an excess of either rat (B1) or bovine (B2) hemoglobin, binding by the labeled hemoglobin was partially inhibited, defined as a reproducible reduction in visible intensity, compared to labeled human hemoglobin alone, following development with the chromogenic substrate. Binding was also partially inhibited by human and bovine globins (data not shown). Although human hemoglobin complexed to human haptoglobin completely inhibited binding of the labeled hemoglobin (B3), heme, sperm whale skeletal muscle myoglobin, human serum albumin (B4), bovine serum albumin (B5), or human transferrin (B6) did not affect binding of the labeled hemoglobin. Thus, H. influenzae specifically binds human hemoglobin and possibly the human hemoglobin-haptoglobin complex since biotinylated hemoglobin binding is blocked by human hemoglobin-haptoglobin complex. Hemoglobin binding by H. influenzae is preferential for human hemoglobin. Partial inhibition of human hemoglobin binding in the presence of rat and bovine hemoglobins suggests that these hemoglobins are also recognized by H. influenzae, although they bind with a lower affinity. These results are similar to those reported for bacterial transferrin receptors which are highly specific for the transferrin of their host species [1]. This

4 246 specificity may be due to differences among species in amino acid sequence or subsequent conformational changes [1,12]. The heme residue appears to be necessary for efficient binding since human and bovine globins only partially inhibited binding. This may be attributable to conformational changes which occur upon loss of the heme moiety [12]. Heme itself does not seem to be directly responsible for binding, since binding of biotinylated human hemoglobin was not inhibited in the presence of heme. In contrast, hemoglobin binding by Neisseria meningitidis was competitively inhibited by heme, as well as bovine hemoglobin and the heme-containing compound catalase, indicating that the heme moiety is recognized [13]. The same investigator has previously reported two heme binding proteins in N. gonorrhoeae; in this case heme binding could be inhibited by hemoglobin and the heme-containing compound equine cytochrome ell 1 [14]. It is possible that the hemoglobin binding property observed in N. meningitidis may in fact represent a 2.5' I I I I I I i I I I I I I I I I 0 Hb B I Binding 4, _- 4, 4, _- 4, -- 4, m Genetic Distance Fig. 3. Dendrogram demonstrating the distribution of the hemoglobin binding phenotype within the species H. influenzae. Strains indicated by a dot (e) are type b strains and the remainder are serologically non-typable strains. Adapted from [16]. m 0 c 13.0 < 2.0" 1.5" , - ' ', ' ', ' ' ', ' ', " ' ', ' ' ' ' ' ' - ' ' ' ' ' ' Hb B (ng) Fig. 2. Whole-cell enzyme-linked hemoglobin binding assay demonstrating saturability of hemoglobin-binding by Haemophilus influenzae cells grown under heme-deplete conditions cfu were incubated with increasing concentrations (0-200 ng ml- l; nm) of biotinylated human hemoglobin (HbB), followed by extensive washes and then incubation with streptavidin-horseradish peroxidase conjugate. After further washes cells were incubated with the chromogenic substrate ABTS, and absorbance at 405 nm determined. Absorbance is a measure of cell-associated Hb B. heme binding property analogous to N. gonorrhoeae. In contrast, our data indicate specific binding of hemoglobin by H. influenzae since binding was completely inhibited by human hemoglobin but not at all by heme or the hemecontaining compound sperm whale myoglobin. In addition, partial inhibition of binding by both human and bovine globins indicated that it is the protein portion of the hemoglobin molecule that is recognized. A whole-cell enzyme-linked hemoglobin binding assay was used to determine whether the hemoglobin binding property demonstrated saturability. Fig. 2 shows that hemoglobin binding was saturable with ng ( mol) of hemoglobin per 10 9 cfu. This saturability pattern is characteristic of a receptor-ligand interaction [15]. To localize the hemoglobin binding property of H. influenzae, outer membrane protein preparations from heme-deplete and heme-replete type b strain HI689 were utilized in the dot blot assay.

5 247 Outer membrane proteins from cells grown under heme-replete conditions did not bind the labeled hemoglobin while outer membrane proteins from cells grown under heme-deplete conditions bound hemoglobin. To determine whether proteolytic digestion of H. influenzae whole cells would affect the hemoglobin binding property, cells were treated with trypsin prior to performing dot blots. Cells grown in heme-deplete media and subsequently incubated with trypsin did not bind biotinylated hemoglobin. These data suggest that hemoglobin binding is mediated via a surface-exposed outer membrane protein or proteins. To determine the distribution of hemoglobin binding within the species, H. influenzae strains previously characterized by multilocus enzyme electrophoresis were investigated [16]. With the exception of a single genetically highly divergent type b strain, human hemoglobin binding was detected by dot blot analysis in all tested type b and non-typable H. influenzae (Fig. 3). In addition, strains representing serotypes a and c-f bound biotinylated human hemoglobin when grown in heme-deplete media (Table 1). These data demonstrate that hemoglobin binding is highly conserved and broadly distributed among both encapsulated and serologically non-typable strains of 1t. influenzae. Other bacterial species were also investigated for hemoglobin binding. Among the Haemophilus species, only H. influenzae and H. haemolyticus bound human hemoglobin (Table 1). H. influenzae and H. haemolyticus are the only two Haemophilus species tested which are unable to synthesize porphyrin. Binding of human hemoglobin was not detected in H. parainfluenzae, H. Table 1 Species distribution of binding of human hemoglobin Bacterial species H. influenzae (serotypes a < f) + H. influenzae nt + H. parainfluenzae H. haemolyticus + H. parahaemolyticus H. aphrophilus E. coli Hb binding parahaemolyticus, 11. aphrophilus, Or E. coil strain HBI01 under iron-deplete conditions. Both H. influenzae and H. haemolyticus bound human hemoglobin when grown under iron-deplete conditions, identical to those used for growth of the other haemophili, as well as following growth in heme-deplete conditions. We have identified a saturable, human hemoglobin specific binding property in H. influenzae, and possibly H. haemolyticus, which is apparently mediated via a surface-exposed outer membrane protein. Binding of hemoglobin may represent the first step in the uptake of heme from hemoglobin. Work is in progress to further characterize the molecular structures mediating hemoglobin binding by H. influenzae, to determine their role in acquisition of heme from hemoglobin and to determine their potential as vaccine candidates. Acknowledgements This work was supported by Public Health Service Grant AI29611 from the National Institute of Allergy and Infectious Diseases and by the Kulicke Foundation. Joseph V. Frangipane and John J. LiPuma are thanked for their assistance in the preparation of the manuscript. References 1 Williams, P. and Griffiths, E. (1992) Bacterial transferrin receptors - structure, function and contribution to virulence. Med. Microbiol. Immunol. 181, Bezkorovainy, A. (1987) Iron proteins. In: Iron and Infection: Molecular, Physiological and Clinical Aspects (Bullen, J.J. and Griffiths, E., Eds.), pp John Wiley and Sons, New York, NY. 3 Morton, D.J. and Williams, P. (1989) Utilization of transferrin-bound iron by Haemophilus species of human and porcine origins. FEMS Microbiol. Lett. 65, Morton, D.J. and Williams, P. (1990) Siderophore-independent acquisition of transferrin-bound iron by Haemophilus influenzae. J. Gen. Microbiol. 136, Turk, D.C. (1984) The pathogenicity of Haemophilus influenzae. J. Med. Microbiol. 18, Pidcock, K.A., Wooten, J.A., Daley, B.A. and Stull, T.L. (1988) Iron acquisition of Haemophilus influenzae. Infect. Immun. 56,

6 248 7 Stull, T.L. (1987) Protein sources of heme for Haemophilus influenzae. Infect. Immun. 55, Williams, P., Morton, D.J., Towner, K.J., Stevenson, P. and Griffiths, E. (1990) Utilization of enterobactin and other exogenous iron sources by Haemophilus influenzae, H. parainfluenzae, and H. paraphrophilus. J. Gen. Microbiol. 136, Morton, D.J., Musser, J.M. and Stull, T.L. (1993) Expression of the Haemophilus influenzae transferrin receptor is repressible by hemin but not by elemental iron alone. Infect. Immun. 61, Shurin, S.B., Anderson, P., Zellinger, J. and Rathbun, R.K. (1986) Pathophysiology of hemolysis in infections with Haemophilus influenzae type b. J. Clin. Invest. 77, Stull, T.L., Mack, K., Hass, J.E., Smith, J. and Smith, A.L. (1985) A comparison of techniques for isolation of the outer membrane proteins of Haemophilus influenzae. Anal. Biochem. 150, Bunn, H.F. and Forget, B.G. (1986) Hemoglobin: Molecular, Genetic and Clinical Aspects. WB Saunders, Philadelphia, PA. 13 Lee, B.C. and Hill, P. (1992) Identification of an outermembrane haemoglobin-binding protein in Neisseria meningitidis. J. Gen. Microbiol. 138, Lee, B.C. (1992) Isolation of haemin-binding proteins of Neisseria meningitidis. J. Med. Microbiol. 36, Kahn, C.R. (1976) Membrane receptors for hormones and neurotransmitters. J. Cell Biol. 70, Musser, J.M., Barenkamp, S.J., Granoff, D.M. and Selander, R.K. (1986) Genetic relationships of serologically nontypeable and serotype b strains of Haemophilus influenzae. Infect. Immun. 52,